本文整理汇总了Python中lshash.LSHash.load_index方法的典型用法代码示例。如果您正苦于以下问题：Python LSHash.load_index方法的具体用法？Python LSHash.load_index怎么用？Python LSHash.load_index使用的例子？那么, 这里精选的方法代码示例或许可以为您提供帮助。您也可以进一步了解该方法所在类lshash.LSHash的用法示例。
在下文中一共展示了LSHash.load_index方法的2个代码示例，这些例子默认根据受欢迎程度排序。您可以为喜欢或者感觉有用的代码点赞，您的评价将有助于系统推荐出更棒的Python代码示例。
示例1: init
# 需要导入模块: from lshash import LSHash [as 别名]
# 或者: from lshash.LSHash import load_index [as 别名]
def init():
parser = argparse.ArgumentParser(description = 'Tools for hamming distance-based image retrieval by cuda')
parser.add_argument('-f', help = 'The filename of image raw features (SIFT).')
parser.add_argument('-v', default = 'fvecs', help = 'The format of image raw features.')
parser.add_argument('-s', default = 'dict', help = 'The method of indexing storage.')
parser.add_argument('-d', default = '128', help = 'Dimensions of raw image feature.')
parser.add_argument('-o', default = '0', help = 'Offset of accessing raw image features.')
parser.add_argument('-n', default = '1', help = 'Number of raw image features to read.')
parser.add_argument('-i', default = 'n', help = 'Whether to perform indexing step.')
parser.add_argument('-e', help = 'The dirname of indexing folder.')
parser.add_argument('-k', default = '10', help = 'Number of retrieved images.')
parser.add_argument('-r', default = '32', help = 'Number of dimensions randomly sampled.')
parser.add_argument('-c', default = 'n', help = 'Whether to perform compressing step.')
parser.add_argument('-q', default = 'n', help = 'Whether to sequentially sampling.')
parser.add_argument('-p', default = 'n', help = 'Whether to perform querying in compressed domain.')
parser.add_argument('-g', default = 'y', help = 'GPU mode. default is "yes".')
parser.add_argument('-l', default = 'n', help = 'VLQ base64 mode. Load VLQ base64 encoding compressed dict.')
parser.add_argument('-b', default = '1', help = 'Expanding level of search buckets.')
parser.add_argument('-t', default = 'int32', help = 'FastDict type (int32, int8, string).')
args = parser.parse_args()
d = int(args.d)
nuse = int(args.n)
off = int(args.o)
random_dims = int(args.r)
random_sampling = True
if args.q == 'y':
random_sampling = False
lsh = LSHash(64, d, random_sampling, args.t, random_dims, 1, storage_config = args.s, matrices_filename = 'project_plane.npz')
np_feature_vecs = load_features(args.f, args.v, nuse, d, off)
if args.c != 'y' and args.i != 'y' and args.e != None and args.s == 'random':
if args.p == 'y':
print "loading compressed index."
lsh.load_compress_index(args.e, (args.l == 'y'))
print "loading done."
else:
print "loading index."
lsh.load_index(args.e)
print "loading done."
print "indexing done. Ready for querying."
return (lsh, np_feature_vecs, args)示例2: main
# 需要导入模块: from lshash import LSHash [as 别名]
# 或者: from lshash.LSHash import load_index [as 别名]
def main():
parser = argparse.ArgumentParser(description = 'Tools for hamming distance-based image retrieval by cuda')
parser.add_argument('-f', help = 'The filename of image raw features (SIFT).')
parser.add_argument('-v', default = 'fvecs', help = 'The format of image raw features.')
parser.add_argument('-s', default = 'dict', help = 'The method of indexing storage.')
parser.add_argument('-d', default = '128', help = 'Dimensions of raw image feature.')
parser.add_argument('-o', default = '0', help = 'Offset of accessing raw image features.')
parser.add_argument('-n', default = '1', help = 'Number of raw image features to read.')
parser.add_argument('-i', default = 'n', help = 'Whether to perform indexing step.')
parser.add_argument('-e', help = 'The dirname of indexing folder.')
parser.add_argument('-k', default = '10', help = 'Number of retrieved images.')
parser.add_argument('-r', default = '32', help = 'Number of dimensions randomly sampled.')
parser.add_argument('-c', default = 'n', help = 'Whether to perform compressing step.')
parser.add_argument('-q', default = 'n', help = 'Whether to sequentially sampling.')
parser.add_argument('-p', default = 'n', help = 'Whether to perform querying in compressed domain.')
parser.add_argument('-g', default = 'y', help = 'GPU mode. default is "yes".')
parser.add_argument('-l', default = 'n', help = 'VLQ base64 mode. Load VLQ base64 encoding compressed dict.')
parser.add_argument('-b', default = '1', help = 'Expanding level of search buckets.')
parser.add_argument('-t', default = 'int32', help = 'FastDict type (int32, int8, string).')
parser.add_argument('-u', default = 'local', help = 'CUDA client type (local, net).')
parser.add_argument('-host', default = 'localhost', help = 'CUDA server address.')
args = parser.parse_args()
d = int(args.d)
nuse = int(args.n)
off = int(args.o)
random_dims = int(args.r)
random_sampling = True
if args.q == 'y':
random_sampling = False
lsh = LSHash(64, d, random_sampling, args.t, args.u, args.host, random_dims, 1, storage_config = args.s, matrices_filename = 'project_plane.npz')
np_feature_vecs = load_features(args.f, args.v, nuse, d, lsh, args.e, off, args.i)
if args.c == 'y':
if args.e != None and args.s == 'random':
lsh.load_index(args.e)
print "compressing index..."
lsh.compress_index(args.e)
print "compressing done."
else:
print "Please specify generated indexing file."
sys.exit(0)
if args.c != 'y' and args.i != 'y' and args.e != None and args.s == 'random':
if args.p == 'y':
print "loading compressed index."
lsh.load_compress_index(args.e, (args.l == 'y'))
print "loading done."
else:
print "loading index."
lsh.load_index(args.e)
print "loading done."
if args.p != 'y':
retrived = lsh.query(np_feature_vecs[1], num_results = int(args.k), expand_level = int(args.b), distance_func = 'hamming')
else:
retrived = lsh.query_in_compressed_domain(np_feature_vecs[1], num_results = int(args.k), expand_level = int(args.b), distance_func = 'hamming', gpu_mode = args.g, vlq_mode = args.l)
print retrived